Generative AI improves emergency message translations but needs human checks
Generative AI translations in high-stakes emergency messaging
Computation and Language
Summary
Emergency messages like earthquake warnings must be clear in many languages, because mistakes can be very dangerous. The authors tested AI translating such messages from English to Chinese and Spanish using special instructions that helped make the translations easier to understand and act upon. However, translators still didn’t fully trust these AI outputs and emphasized the need for human review to find errors and take responsibility for the messages. The study shows AI can speed up initial translations but cannot replace humans when the stakes are high.
What this means in practice
- •For emergency management teams: Generate initial translations of urgent safety messages more quickly in multiple languages to enable faster revision and authorization.
- •For local government communication offices: Use guided AI translation to expand the range of languages for public safety instructions while ensuring human responsibility for accuracy.
Authors
Nune Ayvazan, Anthony Pym, Yu Hao
Abstract
Emergency messaging such as extreme-weather reports and earthquake instructions can involve high stakes, to the extent that translation errors can lead to tragic consequences. The use of machine translation or generative artificial intelligence might therefore not be recommended. On the other hand, time savings in the initial translation can allow greater investments of resources in revision and authorization processes, as well as a wider range of target languages. An experiment with generative AI translations of an earthquake instruction text from English into Chinese and Spanish shows that use of discourse-specific prompts can considerably improve understandability and actionability, although the translations may still not be trusted by translators. Human revision is still required, not only to detect errors but also because of the ethical need for someone to take responsibility for any errors or delays in such messaging.